Committee Members Information

• Ph.D. Advisor: Sigurjon Jonsson
• External Examiner: Peter Christopher LaFemina
• Committee Chair: Abdulkader M. Afifi
• 4th Committee Member: Paul Martin Mai

Abstract

Iceland is one of the few places where ridge–transform interactions can be observed above sea level. This accessibility allows us to study these transtensional settings with space geodetic techniques. These direct observations are essential for the understanding of the present-day kinematics of such settings and for regional geohazards assessments. This thesis investigates plate boundary deformation processes in Iceland by integrating space geodetic and seismic observations across two transtensional settings: the Tjörnes Fracture Zone in the north and the Reykjanes Peninsula in the southwest. The Tjörnes Fracture Zone in northern Iceland is a complex transform system capable of producing magnitude ~7 earthquakes that links offshore and onshore ridge segments. Within this zone, the Húsavík–Flatey Fault, a strike-slip fault with a small opening component, offers a unique opportunity for geodetic studies as it goes onshore for about 25 km. Using more than two decades of continuous and campaign GNSS measurements, I produced the most up-to-date velocity field for North Iceland and applied a backslip model to reproduce interseismic deformation. My results indicate that the Húsavík-Flatey Fault accommodates ~35% of the relative plate motion in the Tjörnes Fracture Zone, with the rest taken up by the Grímsey Oblique Rift. At a slip rate of ~7 mm/yr and a uniform locking depth of ~7 km, these results imply that since the last major earthquakes in 1872, the accumulated slip deficit is equivalent to a Mw~6.9 earthquake. I then analyze the three mainshocks (Mw 5.4, 5.7, and 6.0) of the June 2020 earthquake swarm in the Tjörnes Fracture Zone by combining GNSS coseismic offsets and local and regional seismic waveforms through Bayesian inference. The results indicate strike-slip faulting for the first two mainshocks: the first on a branch of the Húsavík-Flatey Fault and the second on a conjugate fault. The third mainshock corresponds to a normal-faulting earthquake on the eastern flank of the Eyjafjarðaráll Rift. The depths of all events are between ~3–6 km, consistent with seismotectonic studies in the region. There is a systematic underestimation of the magnitudes from the three mainshocks (Mw 5.3, 5.6, and 5.9) compared to magnitudes obtained from teleseismic data. The geodetic magnitudes (Mw5.52, 5.71, and 5.96) obtained from a model with fixed depth and geometry would indicate an aseismic moment release of 40%. However, the actual aseismic fraction is likely lower.
Finally, I use the recent volcanic unrest in the Reykjanes Peninsula, which caused fracturing and reactivation of structures, to explore the potential of deep learning for automated fracture segmentation in wrapped interferograms. At present, the small dataset and high noise levels limit model performance. However, the trained model (IoU~40%) successfully maps major structures and identifies fractures absent from the labels, demonstrating the potential of this approach for future applications. Together, these studies improve our understanding of the plate boundary deformation processes in Iceland, from long-term interseismic strain accumulation in the Tjörnes Fracture Zone to the role of earthquake swarms in releasing accumulated strain. In addition, I provide a methodological advancement to improve mapping efficiency during rapidly evolving volcanic crises. Ultimately, the results obtained here contribute to seismic and volcanic hazard assessment in Iceland, while the insights gained can be extrapolated to inaccessible submarine systems, including those in the Red Sea.